White matter hyperintensities in progranulin-associated frontotemporal dementia: A longitudinal GENFI study

Abstract

Frontotemporal dementia (FTD) is a heterogeneous group of neurodegenerative disorders with both sporadic and genetic forms. Mutations in the progranulin gene (GRN) are a common cause of genetic FTD, causing either a behavioural presentation or, less commonly, language impairment. Presence on T2-weighted images of white matter hyperintensities (WMH) has been previously shown to be more commonly associated with GRN mutations rather than other forms of FTD. The aim of the current study was to investigate the longitudinal change in WMH and the associations of WMH burden with grey matter (GM) loss, markers of neurodegeneration and cognitive function in GRN mutation carriers. 336 participants in the Genetic FTD Initiative (GENFI) study were included in the analysis: 101 presymptomatic and 32 symptomatic GRN mutation carriers, as well as 203 mutation-negative controls. 39 presymptomatic and 12 symptomatic carriers, and 73 controls also had longitudinal data available. Participants underwent MR imaging acquisition including isotropic 1 mm T1-weighted and T2-weighted sequences. WMH were automatically segmented and locally subdivided to enable a more detailed representation of the pathology distribution. Log-transformed WMH volumes were investigated in terms of their global and regional associations with imaging measures (grey matter volumes), biomarker concentrations (plasma neurofilament light chain, NfL, and glial fibrillary acidic protein, GFAP), genetic status (TMEM106B risk genotype) and cognition (tests of executive function). Analyses revealed that WMH load was higher in both symptomatic and presymptomatic groups compared with controls and this load increased over time. In particular, lesions were seen periventricularly in frontal and occipital lobes, progressing to medial layers over time. However, there was variability in the WMH load across GRN mutation carriers - in the symptomatic group 25.0% had none/mild load, 37.5% had medium and 37.5% had a severe load - a difference not fully explained by disease duration. GM atrophy was strongly associated with WMH load both globally and in separate lobes, and increased WMH burden in the frontal, periventricular and medial regions was associated with worse executive function. Furthermore, plasma NfL and to a lesser extent GFAP concentrations were seen to be associated with increased lesion burden. Lastly, the presence of the homozygous TMEM106B rs1990622 TT risk genotypic status was associated with an increased accrual of WMH per year. In summary, WMH occur in GRN mutation carriers and accumulate over time, but are variable in their severity. They are associated with increased GM atrophy and executive dysfunction. Furthermore, their presence is associated with markers of WM damage (NfL) and astrocytosis (GFAP), whilst their accrual is modified by TMEM106B genetic status. WMH load may represent a target marker for trials of disease modifying therapies in individual patients but the variability across the GRN population would prevent use of such markers as a global outcome measure across all participants in a trial.

Keywords: Dementia; Frontotemporal dementia; Progranulin; White matter hyperintensities.

Fig. 1

Top row: Marginal average of white matter hyperintensity (WMH) burden in the individual lobes and layers after correction for age, gender, scanner and TIV in controls, presymptomatic and symptomatic GRN mutation carriers. The bottom row shows a guide to the figures [left, lobar subdivision; right, layer subdivision]. The colour bar represents the average WMH load (increased = red, less = light yellow). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

3D representation of the main tracts passing through the average white matter lesion maps of the GRN mutation carriers: in orange the tracts affected by the presence of lesions, and in green the tracts that do not go through lesions. The average lesion location is coloured in red. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Significant associations between cross-sectional grey matter (GM) volume (Tot = total, F = frontal, P = parietal, O = occipital, T = temporal) and white matter hyperintensity (WMH) burden (Tot = total, F = frontal, P = parietal, O = occipital, T = temporal, 1 = periventricular layer, Med = medial layers, 4 = peripheral layer) within the GRN population. Significance is defined at a p-value threshold of 0.05 in the linear regression between GM volume and log transformed WMH volume after correction for age, gender, total intracranial volume and scanner type.